Transformer distribution type distributed resource dynamic aggregation method, device, equipment, chip and medium

Abstract

The present disclosure relates to the technical field of computers, in particular to a transformer area distributed resource dynamic aggregation method, device, equipment, chip and medium, the transformer area distributed resource dynamic aggregation method comprises: obtaining a sample of the distributed resource, determining a degree matrix and an adjacency matrix of the distributed resource based on the sample; determining a standardized Laplacian matrix of the sample based on the degree matrix and the adjacency matrix; determining a cluster matrix of the sample based on the standardized Laplacian matrix; clustering the cluster matrix using a genetic algorithm to obtain a plurality of large-area aggregation clusters, each large-area aggregation cluster having a cluster center; taking each cluster center as a large-area control node of the transformer area, and realizing transformer area distributed resource dynamic aggregation based on the large-area control node, thereby realizing dynamic aggregation of transformer area distributed resources through the large-area control node to improve aggregation efficiency and enhance the regulation and control capability of the power distribution network.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, specifically to a method, apparatus, device, chip, and medium for dynamic aggregation of distributed resources in a distribution area. Background Technology

[0002] The large-scale construction of intelligent integrated terminal projects in distribution areas has provided effective technical means for realizing the decentralized, dynamic aggregation, and collaborative control of massive distributed resources. However, the following problems still exist in the key technologies for dynamic aggregation and control of distributed resources in multiple distribution areas: (1) The characteristics and operational regulation capabilities of distributed resources are unclear. (2) It is difficult to access and configure distributed resources. (3) Dynamic aggregation of distributed resources is difficult to achieve. (4) There is a lack of cluster control methods for distributed resources in multiple distribution areas. Among them, due to the characteristics of massive, heterogeneous, and uncertain distributed resources, simple aggregation cannot meet the needs of power grid frequency regulation, voltage regulation, peak regulation, inertial support, etc. Dynamic aggregation must be carried out based on the types, capacity, and complementary characteristics of distributed resources. Therefore, in order to solve the problems of distributed resource access, aggregation, and control, it is urgent to propose a dynamic aggregation algorithm for distributed resources in multiple distribution areas based on cloud-edge-device collaboration to improve the control capability of new distribution networks for various types of distributed resources. Summary of the Invention

[0003] To address the problems in related technologies, this disclosure provides a method, apparatus, device, chip, and medium for dynamic aggregation of distributed resources in a distribution area.

[0004] In a first aspect, this disclosure provides a method for dynamic aggregation of distributed resources in a distribution area, including:

[0005] Obtain a sample of the distributed resource, and determine the degree matrix and adjacency matrix of the distributed resource based on the sample;

[0006] The normalized Laplacian matrix of the sample is determined based on the degree matrix and the adjacency matrix.

[0007] The clustering matrix of the samples is determined based on the standardized Laplacian matrix;

[0008] The clustering matrix is ​​clustered using a genetic algorithm to obtain multiple large-area clusters, each of which has a cluster center.

[0009] Each cluster center serves as the regional control node for the transformer area, and dynamic aggregation of distributed resources in the transformer area is achieved based on the regional control node.

[0010] According to embodiments of this disclosure, the step of using a genetic algorithm to cluster the clustering matrix to obtain multiple large-area clusters includes:

[0011] Optimize the clustering matrix using the first objective function and the second objective function of the genetic algorithm to obtain a preliminary clustering result;

[0012] Optimize the number of clustering clusters for the preliminary clustering result using the third objective function of the genetic algorithm to obtain the multiple large - area aggregation clusters.

[0013] According to an embodiment of the present disclosure, the optimizing the clustering matrix using the first objective function and the second objective function of the genetic algorithm to obtain a preliminary clustering result includes:

[0014] Randomly select an initial clustering center p, and allocate the element ɑ of the clustering matrix to the corresponding cluster based on the minimum Euclidean distance criterion, where i is a positive integer, 0 < i ≤ N, and N is the total number of elements in the clustering matrix; Randomly select an initial clustering center p, and allocate the element ɑ of the clustering matrix to the corresponding cluster based on the minimum Euclidean distance criterion, where i is a positive integer, 0 < i ≤ N, and N is the total number of elements in the clustering matrix;

[0015] Recalculate the new clustering center p of the cluster to which the element ɑ is allocated based on the first objective function; i Recalculate the new clustering center p of the cluster to which the element ɑ is allocated based on the first objective function; new ;

[0016] Re - cluster the current cluster based on the new clustering center p to obtain a new current cluster C; new Re - cluster the current cluster based on the new clustering center p to obtain a new current cluster C; new ;

[0017] Verify the distance between the new current cluster C and other clusters based on the second objective function to ensure that the element ɑ is allocated to the correct cluster. new Verify the distance between the new current cluster C and other clusters based on the second objective function to ensure that the element ɑ is allocated to the correct cluster. i Verify the distance between the new current cluster C and other clusters based on the second objective function to ensure that the element ɑ is allocated to the correct cluster.

[0018] According to an embodiment of the present disclosure, the first objective function is a cluster effectiveness objective function based on the within - cluster cohesion index, and the within - cluster cohesion index is determined by the mean square deviation E of the within - cluster distance,

[0019] where E j is the mean square deviation of the within - cluster distance of the j - th element in the current cluster, j is a positive integer, 1 ≤ j ≤ M, and M is the total number of elements in the current cluster.

[0020] According to an embodiment of the present disclosure, the recalculating the new clustering center of the cluster to which the element ɑ is allocated based on the first objective function includes: i According to an embodiment of the present disclosure, the recalculating the new clustering center of the cluster to which the element ɑ is allocated based on the first objective function includes:

[0021] Recalculate the mean square deviation of the within - cluster distance of each element in the cluster to which the element ɑ is allocated based on the first objective function; i Recalculate the mean square deviation of the within - cluster distance of each element in the cluster to which the element ɑ is allocated based on the first objective function;

[0022] Determine the element with the minimum mean square deviation of the within - cluster distance as the new clustering center of the cluster.​

[0023] According to embodiments of this disclosure, the second objective function is a cluster effectiveness objective function based on an inter-cluster separation index, wherein the inter-cluster separation index is determined by the inter-cluster distance I, and I = max‖C new -C m ‖, where m is a positive integer, and m is less than or equal to the total number of clusters in the clustering matrix, ‖C new -C m || is the calculation of the new current cluster C new In and cluster C m The difference between the corresponding elements in the middle.

[0024] According to embodiments of this disclosure, the verification of the new current cluster C based on the second objective function... new Distance from other clusters to ensure element α i Being assigned to the correct cluster, including:

[0025] Calculate the new current cluster C new The element α is determined when the inter-cluster distance I with other clusters is greater than a first threshold. i It was assigned to the correct cluster.

[0026] According to embodiments of this disclosure, the third objective function is an inter-cluster correlation objective function, determined by the inter-cluster covariance coefficient, which is:

[0027]

[0028] Where a is the cluster center of the first cluster A, and b is the cluster center of the second cluster B. It is the average value of all elements in the first cluster A. It is the average value of each element in the second cluster B.

[0029] According to embodiments of this disclosure, the step of optimizing the number of clusters in the preliminary clustering results using a third objective function to obtain multiple large-area clusters includes:

[0030] Calculate the inter-cluster covariance coefficient between each cluster, and merge clusters whose inter-cluster covariance coefficient is less than a second threshold;

[0031] Calculate the inter-cluster covariance coefficients again after merging, until the inter-cluster covariance coefficients between any two clusters are greater than or equal to the second threshold.

[0032] The final clusters are determined to be the regional aggregation clusters.

[0033] According to embodiments of this disclosure, the dynamic aggregation of distributed resources in a distribution area based on the regional control node includes:

[0034] Frequency modulation aggregation and / or peak modulation aggregation of distributed resources in the transformer area are realized based on the regional control node.

[0035] According to embodiments of this disclosure, the regional control node is based on a response time T. respond Response duration τ duration and / or responsive frequency ω respond Dynamically aggregate the distributed resources of the aforementioned transformer area;

[0036] Wherein, the response time T respond The response duration τ is the maximum time required for a resource to fully reach the output level required by the instruction after receiving it. duration After receiving a resource response command, the response must maintain the output state required by the command for the shortest possible time, wherein the responsive frequency ω respond This represents the number of times a resource can be accessed within a given period of time.

[0037] According to embodiments of this disclosure, the regional control node dynamically aggregates the distributed resources based on load frequency regulation performance indicators and / or load peak regulation indicators;

[0038] Wherein, the load frequency regulation performance index Fn = (1-t) respond )ω respond ,

[0039] t respond =T respond / max(T 1，respond T 2，respond ,…,T N，respond );

[0040] The load peak shaving index Pn = (1-t) respond )τ duration .

[0041] According to embodiments of this disclosure, the distributed resources participating in frequency modulation need to meet the following constraints:

[0042] T n，respond ≤T F respond

[0043] τ n，duration ≥τ F duration

[0044] ω n，respond ≥ω F respond ;

[0045] Among them, T F respond τ is the upper limit of the acceptable response time for FM service.F duration For the duration of the response of frequency modulation service, ω F respond For the maximum responsive frequency; and / or

[0046] Distributed resources participating in peak shaving must meet the following constraints:

[0047] T n，respond ≤T P respond

[0048] τ n，duration ≥τ P duration ;

[0049] Among them, T P respond and τ P duration These are the maximum allowed response time and the minimum allowed duration for peak shaving services, respectively.

[0050] According to embodiments of this disclosure, obtaining a sample of the distributed resource and determining the degree matrix D of the distributed resource based on the sample includes:

[0051] The load characteristic parameters of the distributed resources are obtained as the sample;

[0052] The characteristic parameters of the distributed resource are determined based on the Euclidean distance of the load characteristic parameters;

[0053] The degree matrix is ​​constructed based on the feature parameters.

[0054] According to embodiments of this disclosure, determining the adjacency matrix of the distributed resource includes:

[0055] The adjacency matrix is ​​constructed based on the Gaussian kernel function.

[0056] According to embodiments of this disclosure, determining the clustering matrix of the samples based on the standardized Laplacian matrix includes:

[0057] The eigenvalues ​​and eigenvectors of the standardized Laplacian matrix;

[0058] The clustering matrix is ​​constructed by using the eigenvectors corresponding to the top k largest eigenvalues ​​as row vectors.

[0059] Secondly, this disclosure provides a dynamic aggregation device for distributed resources in a distribution area, comprising:

[0060] The first determining module is configured to obtain a sample of the distributed resource and determine the degree matrix and adjacency matrix of the distributed resource based on the sample.

[0061] A second determination module, configured to determine a normalized Laplacian matrix of the sample based on the degree matrix and the adjacency matrix;

[0062] A third determination module, configured to determine a clustering matrix of the sample based on the normalized Laplacian matrix;

[0063] A clustering module, configured to cluster the clustering matrix by using a genetic algorithm to obtain multiple large - area aggregation clusters, and each of the large - area aggregation clusters has a clustering center;

[0064] An aggregation module, configured to use each clustering center as a large - area control node of the台区, and implement dynamic aggregation of distributed resources in the台区 based on the large - area control node.

[0065] According to an embodiment of the present disclosure, the clustering the clustering matrix by using a genetic algorithm to obtain multiple large - area aggregation clusters includes:

[0066] Optimizing the clustering matrix by using a first objective function and a second objective function of the genetic algorithm to obtain a preliminary clustering result;

[0067] Optimizing the number of clustering clusters of the preliminary clustering result by using a third objective function of the genetic algorithm to obtain the multiple large - area aggregation clusters.

[0068] According to an embodiment of the present disclosure, the optimizing the clustering matrix by using a first objective function and a second objective function of the genetic algorithm to obtain a preliminary clustering result includes:

[0069] Randomly selecting an initial clustering center p, and allocating an element ɑ of the clustering matrix to a corresponding cluster based on a minimum Euclidean distance criterion, where i is a positive integer, 0 < i ≤ N, and N is the total number of elements in the clustering matrix; i where i is a positive integer, 0 < i ≤ N, and N is the total number of elements in the clustering matrix;

[0070] Recalculating a new clustering center p of the cluster to which the element ɑ is allocated based on the first objective function; i of the cluster to which the element ɑ is allocated; new ;

[0071] Re - clustering the current cluster based on the new clustering center p to obtain a new current cluster C; new to obtain a new current cluster C; new ;

[0072] Verifying the distance between the new current cluster C and other clusters based on the second objective function to ensure that the element ɑ is allocated to the correct cluster. new is allocated to the correct cluster. i is allocated to the correct cluster.

[0073] According to embodiments of this disclosure, the first objective function is a cluster effectiveness objective function based on an intra-cluster cohesion index, wherein the intra-cluster cohesion index is determined by the mean square deviation E of the intra-cluster distance.

[0074] Among them, E j Let M be the mean squared deviation of the intra-cluster distance of the j-th element in the current cluster, where j is a positive integer, 1≤j≤M, and M is the total number of elements in the current cluster.

[0075] According to embodiments of this disclosure, the element α is recalculated and allocated based on the first objective function. i The new cluster centers of the clusters include:

[0076] The element α was recalculated and redistributed based on the first objective function. i The mean square deviation of the intra-cluster distances of elements in a cluster;

[0077] The element with the smallest mean square deviation of the intra-cluster distance is determined as the new cluster center of the cluster.

[0078] According to embodiments of this disclosure, the second objective function is a cluster effectiveness objective function based on an inter-cluster separation index, wherein the inter-cluster separation index is determined by the inter-cluster distance I, and I = max‖C new -C m ‖, where m is a positive integer, and m is less than or equal to the total number of clusters in the clustering matrix, ‖C new -C m || is the calculation of the new current cluster C new In and cluster C m The difference between the corresponding elements in the middle.

[0079] According to embodiments of this disclosure, the verification of the new current cluster C based on the second objective function... new Distance from other clusters to ensure element α i Being assigned to the correct cluster, including:

[0080] Calculate the new current cluster C new The element α is determined when the inter-cluster distance I with other clusters is greater than a first threshold. i It was assigned to the correct cluster.

[0081] According to embodiments of this disclosure, the third objective function is an inter-cluster correlation objective function, determined by the inter-cluster covariance coefficient, which is:

[0082]

[0083] Where a is the cluster center of the first cluster A, and b is the cluster center of the second cluster B. It is the average value of all elements in the first cluster A. It is the average value of each element in the second cluster B.

[0084] According to embodiments of this disclosure, the step of optimizing the number of clusters in the preliminary clustering results using a third objective function to obtain multiple large-area clusters includes:

[0085] Calculate the inter-cluster covariance coefficient between each cluster, and merge clusters whose inter-cluster covariance coefficient is less than a second threshold;

[0086] Calculate the inter-cluster covariance coefficients again after merging, until the inter-cluster covariance coefficients between any two clusters are greater than or equal to the second threshold.

[0087] The final clusters are determined to be the regional aggregation clusters.

[0088] According to embodiments of this disclosure, the dynamic aggregation of distributed resources in a distribution area based on the regional control node includes:

[0089] Frequency modulation aggregation and / or peak modulation aggregation of distributed resources in the transformer area are realized based on the regional control node.

[0090] According to embodiments of this disclosure, the regional control node is based on a response time T. respond Response duration τ duration and / or responsive frequency ω respond Dynamically aggregate the distributed resources of the aforementioned transformer area;

[0091] Wherein, the response time T respond The response duration τ is the maximum time required for a resource to fully reach the output level required by the instruction after receiving it. duration After receiving a resource response command, the response must maintain the output state required by the command for the shortest possible time, wherein the responsive frequency ω respond This represents the number of times a resource can be accessed within a given period of time.

[0092] According to embodiments of this disclosure, the regional control node dynamically aggregates the distributed resources based on load frequency regulation performance indicators and / or load peak regulation indicators;

[0093] Wherein, the load frequency regulation performance index Fn = (1-t) respond )ω respond ,

[0094] t respond =T respond / max(T 1，respond T 2，respond ,…,T N，respond );

[0095] The load peak shaving index Pn = (1-t) respond )τ duration .

[0096] According to embodiments of this disclosure, the distributed resources participating in frequency modulation need to meet the following constraints:

[0097] T n，respond ≤T F respond

[0098] τ n，duration ≥τ F duration

[0099] ω n，respond ≥ω F respond ;

[0100] Among them, T F respond τ is the upper limit of the acceptable response time for FM service. F duration For the duration of the response of frequency modulation service, ω F respond For the maximum responsive frequency; and / or

[0101] Distributed resources participating in peak shaving must meet the following constraints:

[0102] T n，respond ≤T P respond

[0103] τ n，duration ≥τ P duration ;

[0104] Among them, T P respond and τ P duration These are the maximum allowed response time and the minimum allowed duration for peak shaving services, respectively.

[0105] According to embodiments of this disclosure, obtaining a sample of the distributed resource and determining the degree matrix D of the distributed resource based on the sample includes:

[0106] The load characteristic parameters of the distributed resources are obtained as the sample;

[0107] The characteristic parameters of the distributed resource are determined based on the Euclidean distance of the load characteristic parameters;

[0108] The degree matrix is ​​constructed based on the feature parameters.

[0109] According to embodiments of this disclosure, determining the adjacency matrix of the distributed resource includes:

[0110] The adjacency matrix is ​​constructed based on the Gaussian kernel function.

[0111] According to embodiments of this disclosure, determining the clustering matrix of the samples based on the standardized Laplacian matrix includes:

[0112] The eigenvalues ​​and eigenvectors of the standardized Laplacian matrix;

[0113] The clustering matrix is ​​constructed by using the eigenvectors corresponding to the top k largest eigenvalues ​​as row vectors.

[0114] Thirdly, embodiments of this disclosure provide an electronic device including a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method as described in any one of the first aspects.

[0115] Fourthly, embodiments of this disclosure provide a chip that includes the apparatus described in any one of the second aspects.

[0116] Fifthly, this disclosure provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the method as described in any one of the first aspects.

[0117] According to the technical solution provided in the embodiments of this disclosure, the method for dynamic aggregation of distributed resources in a distribution area includes: acquiring samples of the distributed resources; determining the degree matrix and adjacency matrix of the distributed resources based on the samples; determining the normalized Laplace matrix of the samples based on the degree matrix and adjacency matrix; determining the clustering matrix of the samples based on the normalized Laplace matrix; clustering the clustering matrix using a genetic algorithm to obtain multiple large-area clusters, each of the large-area clusters having a cluster center; using each cluster center as a large-area control node for the distribution area; and realizing dynamic aggregation of distributed resources in the distribution area based on the large-area control node. By realizing dynamic aggregation of distributed resources in the distribution area through the large-area control node, intelligent aggregation of distributed resources in multiple distribution areas can be achieved, improving aggregation efficiency and enhancing the control capability of the distribution network.

[0118] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0119] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments, taken in conjunction with the accompanying drawings. In the drawings:

[0120] Figure 1 A flowchart illustrating a method for dynamic aggregation of distributed resources in a transformer substation according to an embodiment of this disclosure is shown.

[0121] Figure 2 The diagram illustrates a specific application scenario of the dynamic aggregation method for distributed resources in a transformer substation according to an embodiment of this disclosure.

[0122] Figure 3 A structural block diagram of a distributed resource dynamic aggregation device according to an embodiment of the present disclosure is shown.

[0123] Figure 4 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0124] Figure 5 A schematic diagram of the structure of a computer system suitable for implementing the method according to embodiments of the present disclosure is shown. Detailed Implementation

[0125] In the following, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings to enable those skilled in the art to readily implement them. Furthermore, for clarity, portions unrelated to the description of exemplary embodiments have been omitted from the drawings.

[0126] In this disclosure, it should be understood that terms such as “comprising” or “having” are intended to indicate the presence of features, figures, steps, behaviors, components, parts or combinations thereof disclosed in this specification, and are not intended to exclude the possibility of the presence or addition of one or more other features, figures, steps, behaviors, components, parts or combinations thereof.

[0127] It should also be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0128] In this disclosure, any operation involving the acquisition of user information or user data, or the display of user information or user data to others, is an operation authorized or confirmed by the user, or actively selected by the user.

[0129] As mentioned above, the following problems still exist in the key technologies for dynamic aggregation and regulation of distributed resources in multiple distribution areas: (1) The characteristics and operational regulation capabilities of distributed resources are unclear. (2) Distributed resource access and configuration are difficult. (3) Dynamic aggregation of distributed resources is difficult to achieve. (4) There is a lack of methods for cluster regulation of distributed resources in multiple distribution areas. Among these, due to the characteristics of distributed resources being massive, heterogeneous, and uncertain, simple aggregation cannot meet the needs of power grid frequency regulation, voltage regulation, peak regulation, inertial support, and other aspects. Dynamic aggregation must be carried out based on the types, capacity, and complementary characteristics of distributed resources.

[0130] In related technologies, some literature considers the complementarity of different microgrids (MGs) and distributed resources, using the energy function of panorama theory to group and aggregate multiple MGs. Some scholars have proposed an electric vehicle (EV) aggregation method based on load aggregators (LAs), considering the uncertainty of EV behavior characteristics and effectively aggregating dispersed EV energy storage resources. Other literature proposes using EV aggregation stations as controllable units, studying active distribution network scheduling strategies with EV aggregation stations participating, demonstrating that active distribution networks can assist in absorbing intermittent distributed resources and reducing power fluctuations through active scheduling of EV aggregation stations. Researchers have combined LAs with MGs, using LA coordinated control as a means to establish a two-layer real-time scheduling model for independent microgrids based on scheduling priorities; simultaneously, others have combined LAs with energy storage, studying a market-oriented optimization decision-making framework for LAs with energy storage devices.

[0131] However, current research focuses on single aggregation in typical long-term system scenarios, lacking research on the uncertainties of a large number of complementary distributed resources. Meanwhile, some scholars have proposed a method to allocate the contribution of distributed power sources to distribution network losses and emission reductions based on their respective responsibilities. This proposed gain allocation method can provide distribution companies with reasonable economic signals and offers theoretical reference for research on the decentralization and dynamic aggregation technologies of distributed resources in multi-regional distribution networks, thus possessing certain reference value.

[0132] In summary, current technologies have made some progress in the distribution and dynamic aggregation of distributed adjustable resources, but there is a lack of research on distributed resource control strategies that consider multiple complementary approaches. In terms of dynamic aggregation technology, the aggregation modeling methods for heterogeneous, massive, and uncertain distributed resources also urgently need further research.

[0133] In view of this, embodiments of this disclosure provide a method for dynamic aggregation of distributed resources in a distribution area, comprising: acquiring samples of the distributed resources; determining the degree matrix and adjacency matrix of the distributed resources based on the samples; determining the normalized Laplace matrix of the samples based on the degree matrix and adjacency matrix; determining the clustering matrix of the samples based on the normalized Laplace matrix; clustering the clustering matrix using a genetic algorithm to obtain multiple large-area clusters, each of the large-area clusters having a cluster center; using each cluster center as a large-area control node for the distribution area, and realizing dynamic aggregation of distributed resources in the distribution area based on the large-area control node. According to the technical solution of embodiments of this disclosure, dynamic aggregation of distributed resources in a distribution area is achieved through large-area control nodes, enabling intelligent aggregation of distributed resources in multiple distribution areas, improving aggregation efficiency, and enhancing the control capability of the distribution network.

[0134] Figure 1 A flowchart illustrating a method for dynamic aggregation of distributed resources in a transformer substation according to an embodiment of this disclosure is shown. Figure 1 As shown, the method for dynamic aggregation of distributed resources in a transformer substation includes the following steps S101-S105:

[0135] In step S101, a sample of the distributed resource is obtained, and the degree matrix and adjacency matrix of the distributed resource are determined based on the sample.

[0136] In step S102, the normalized Laplacian matrix of the sample is determined based on the degree matrix and the adjacency matrix;

[0137] In step S103, the clustering matrix of the samples is determined based on the standardized Laplace matrix;

[0138] In step S104, the clustering matrix is ​​clustered using a genetic algorithm to obtain multiple large-area clusters, each of which has a cluster center;

[0139] In step S105, each cluster center is used as the regional control node of the transformer area, and dynamic aggregation of distributed resources in the transformer area is realized based on the regional control node.

[0140] In this embodiment, the distributed resources in the distribution area may include various new energy sources, such as photovoltaic and wind power, as well as energy storage and load. By calculating the load forecast and the new energy output forecast, and combining them with the grid's operating status, the capacity demand for dynamic aggregation of distributed resources in the distribution area, such as frequency regulation and peak shaving, can be determined. Frequency regulation refers to changing the carrier frequency according to the modulation signal, while peak shaving refers to the power generation department adjusting the generator output to adapt to changes in electricity load. In the distribution network system, during peak electricity consumption periods, the addition of new resources causes changes in system frequency and peak value, all due to the load characteristics of the resources. By clustering the load characteristics of each distributed resource in the distribution area and dynamically aggregating the distributed resources based on the clustering results, frequency regulation and / or peak shaving of the system can be achieved, thereby eliminating the aforementioned changes in system frequency and peak value and ensuring stable grid operation.

[0141] In this embodiment of the disclosure, obtaining a sample of the distributed resources and determining the degree matrix and adjacency matrix of the distributed resources based on the sample may involve first obtaining the number N of distributed resources contained in the transformer area. k Define the sample space Where k and N k All are positive integers, and R represents the set of real numbers, which in this embodiment is k*N k The dataset is represented by a matrix; then, some load characteristic parameters are selected from the sample space to form a diagonal matrix, i.e., the degree matrix D, where the load characteristic parameters include, but are not limited to, the load power, load capacity, and location of the resources. Since the selection of the degree matrix directly affects the clustering effect, resources with similar characteristic parameters can be grouped together, equivalent to a group of virtual clusters, and then appropriate load characteristic parameters are selected to form a diagonal matrix. Since the load power values ​​of different resources within a group of virtual clusters vary greatly, they can first be determined using the formula... Normalize the samples, and then calculate sample x. i and x j Euclidean distance between As feature parameters, they constitute the degree matrix D.

[0142] After constructing the similarity matrix D, the adjacency matrix W also needs to be constructed. The weights of the adjacency matrix represent the similarity between samples. While fully connected methods are commonly used to construct adjacency matrices, these matrices are not sparse. Therefore, this embodiment uses a Gaussian kernel function to construct the adjacency matrix W. Where i, j = 1, 2, ..., N k q i and q j Sample xi and x j At a certain moment in, the value of the load characteristic parameter, such as the load power value, ξ is the bandwidth parameter of the Gaussian kernel function, which can be set according to needs in practice and is not limited here.

[0143] In the embodiment of the present disclosure, the method for determining the normalized Laplacian matrix of the sample based on the degree matrix and the adjacency matrix may be to calculate the Laplacian matrix of the sample according to the formula L = D - W, and then according to the formula L sym = D -1 / 2 LD -1 / 2 calculate the normalized Laplacian matrix of the sample.

[0144] In the embodiment of the present disclosure, the method for determining the clustering matrix of the sample based on the normalized Laplacian matrix may be to calculate the eigenvalues and eigenvectors of the normalized Laplacian matrix, and take the eigenvectors a1, a2,..., a corresponding to the top k largest eigenvalues k as row vectors to form the clustering matrix A of the sample, and the dimension of the clustering matrix A is k * N k .

[0145] According to the technical solution of the embodiment of the present disclosure, the dynamic aggregation of the distributed resources in the substation area is realized through the large area control node, which can realize the intelligent aggregation of the distributed resources in multiple substation areas, improve the aggregation efficiency, and enhance the regulation and control ability of the distribution network.

[0146] In the embodiment of the present disclosure, the method for clustering the clustering matrix by using the genetic algorithm to obtain multiple large area aggregation clusters includes: optimizing the clustering matrix by using the first objective function and the second objective function of the genetic algorithm to obtain a preliminary clustering result; optimizing the number of clustering clusters of the preliminary clustering result by using the third objective function of the genetic algorithm to obtain the multiple large area aggregation clusters.

[0147] Among them, the method for optimizing the clustering matrix by using the first objective function and the second objective function of the genetic algorithm to obtain a preliminary clustering result may include: randomly selecting an initial clustering center p, and based on the minimum Euclidean distance criterion, allocating the element ɑ of the clustering matrix i to the corresponding cluster, where i is a positive integer, 0 < i ≤ N, and N is the total number of elements in the clustering matrix; recalculating the new clustering center p of the cluster to which the element ɑ has been allocated based on the first objective function i ; re-clustering the current cluster based on the new clustering center p new to obtain a new current cluster C new ; verifying the new current cluster C based on the second objective function new ; newDistance from other clusters to ensure element α i It was assigned to the correct cluster.

[0148] Specifically, we can first randomly select an initial cluster center p, and then each element a in the clustering matrix A... i Compare with the cluster center p to calculate element a. i The Euclidean distance d from the cluster center p, Finally, based on the comparison results and the minimum Euclidean distance criterion, element α was selected. i Assign elements to the m-th cluster to achieve element clustering.

[0149] Each time an element is assigned to a cluster, the cluster centers need to be recalculated. When recalculating the cluster centers, it can be assumed that the newly assigned element is the cluster center of the current cluster. Based on the first objective function, the intra-cluster aggregation index of the current cluster is verified, and the new cluster center p of the current cluster is determined according to the intra-cluster aggregation index. new .

[0150] Specifically, the first objective function can be a cluster effectiveness objective function based on the intra-cluster cohesion index, wherein the intra-cluster cohesion index is determined by the mean square deviation E of the intra-cluster distance. Where j is a positive integer, 1≤j≤M, and M is the total number of elements in the current cluster. Based on the above formula, calculate all elements a in the current cluster. j Mean square deviation of intra-cluster distance E j Determine the minimum mean square deviation E. min And determine the minimum mean square deviation E min Corresponding element a j p is the new cluster center of the current cluster. new .

[0151] Subsequently, based on the new cluster center p new Re-cluster the current cluster to obtain a new current cluster C. new Then, the new current cluster C is verified based on the second objective function. new Distance from other clusters to ensure element α i The cluster is assigned to the correct cluster. Specifically, the second objective function can be a cluster effectiveness objective function based on an inter-cluster separation index, which is determined by the inter-cluster distance I, where I = max‖C new -C m ‖, where m is a positive integer, and m is less than or equal to the total number of clusters in the clustering matrix, ‖C new -C m || is the calculation of the new current cluster C new In and cluster C m The difference between the elements at corresponding positions in the cluster. Determine the new current cluster C.new If the inter-cluster distance I to other clusters is greater than a first threshold, then determine the element α. i It was assigned to the correct cluster; otherwise, the element α is considered to be... i If not assigned to the correct cluster, it needs to be reassigned. The value of the first threshold can be set according to actual needs and is not limited here.

[0152] Thus, based on the first and second objective functions, the first clustering optimization of the elements in the clustering matrix has been completed. By reasonably evaluating the fitness of individuals in the initial population, the number of iterations required to achieve accurate partitioning has been optimized. Next, the number of clusters after the first clustering optimization can be further optimized based on the third objective function, while simultaneously generating the optimal cluster centers.

[0153] In this embodiment of the disclosure, the step of optimizing the number of clusters in the preliminary clustering results using a third objective function to obtain multiple large-area aggregated clusters includes: calculating the inter-cluster covariance coefficient between each cluster, and merging clusters whose inter-cluster covariance coefficient is less than a second threshold; calculating the inter-cluster covariance coefficient of each merged cluster again until the inter-cluster covariance coefficient between any two clusters is greater than or equal to the second threshold; and determining that the final obtained clusters are the large-area aggregated clusters.

[0154] To optimize the number of clusters, it is necessary to analyze the correlation between clusters. In this embodiment, the covariance coefficient can be used to represent the correlation between clusters; a positive covariance indicates a positive correlation, and a negative covariance indicates a negative correlation. Specifically, it can be done according to the formula...

[0155] To calculate the covariance coefficient, where a is the cluster center of the first cluster A, and b is the cluster center of the second cluster B. It is the average value of all elements in the first cluster A. This is the average value of each element in the second cluster B. Using the covariance coefficient as the third objective function to measure the correlation between clusters, the correlation coefficients between the clusters after the first clustering optimization are calculated to find clusters with a correlation greater than the second threshold. These clusters are then combined to create a new cluster, thus updating the clusters. The value of the second threshold can be set according to actual needs and is not limited here. After the clusters are updated, each cluster is optimized again based on the first, second, and third objective functions, iterating repeatedly until a preset condition is met. The resulting clusters are the large-area aggregation clusters. The preset condition can be that the number of iterations is greater than the third threshold. The value of the third threshold can be set according to actual needs and is not limited here. Each large-area aggregation cluster has a cluster center, which can serve as the large-area control node for the distribution area. Dynamic aggregation of distributed resources in the distribution area is achieved based on the large-area control node.

[0156] According to the technical solution of this disclosure, the genetic algorithm is optimized by first, second and third objective functions, and the clustering matrix is ​​clustered based on the optimized genetic algorithm to obtain multiple large-area clusters. This can quickly and accurately aggregate distributed resources, obtain the optimal large-area control node, and improve the efficiency of dynamic aggregation of distributed resources in the transformer area.

[0157] In this embodiment of the disclosure, the dynamic aggregation of distributed resources in a distribution area based on the regional control node includes: frequency modulation aggregation and / or peak modulation aggregation of distributed resources in the distribution area based on the regional control node. Specifically, the regional control node can be based on the response time T. respond Response duration τ duration and / or responsive frequency ω respond The distributed resources of the aforementioned distribution area are dynamically aggregated; wherein, the response time T respond The response duration τ is the maximum time required for a resource to fully reach the output level required by the instruction after receiving it. duration After receiving a resource response command, the response must maintain the output state required by the command for the shortest possible time, wherein the responsive frequency ω respond This represents the number of times a resource can be accessed within a given period of time.

[0158] In this embodiment of the disclosure, the regional control node dynamically aggregates the distributed resources based on load frequency regulation performance indicators and / or load peak shaving indicators; wherein, the load frequency regulation performance indicator Fn = (1-t) / ( ... respond )ω respond , t respond =T respond / max(T 1，respond T 2，respond ,…,TN，respond The load peak shaving index Pn = (1-t) respond )τ duration .

[0159] In this embodiment of the disclosure, the distributed resources participating in frequency modulation need to meet the following constraints:

[0160] T n，respond ≤T F respond

[0161] τ n，duration ≥τ F duration

[0162] ω n，respond ≥ω F respond ;

[0163] Among them, T F respond τ is the upper limit of the acceptable response time for FM service. F duration For the duration of the response of frequency modulation service, ω F respond This is the maximum responsive frequency.

[0164] In this embodiment of the disclosure, the distributed resources participating in peak shaving need to meet the following constraints:

[0165] T n，respond ≤T P respond

[0166] τ n，duration ≥τ P duration ;

[0167] Among them, T P respond and τ P duration These are the maximum allowed response time and the minimum allowed duration for peak shaving services, respectively.

[0168] According to the technical solution of the present disclosure, the distributed resources of the transformer area are dynamically aggregated based on the proposed new frequency modulation, peak shaving indicators and / or constraints, thereby optimizing the aggregation scheme and improving the aggregation efficiency.

[0169] Figure 2 A diagram illustrating a specific application scenario of the dynamic aggregation method for distributed resources in a transformer substation according to an embodiment of this disclosure is shown. For example... Figure 2As shown, the distributed resources in the distribution area can be loads, energy storage, wind power, photovoltaics, etc. Each distributed resource can communicate bidirectionally with the dispatch center. The dispatch center is connected to the control bus, and bidirectional communication can also be achieved between the dispatch center and the control bus. The control bus can include Nc regional control nodes, which are determined by the above-mentioned dynamic aggregation method of distributed resources in the distribution area. Based on the regional control nodes, the distributed resources in the distribution area can be dynamically aggregated, thereby providing stable and high-quality electricity resources reasonably for electricity users such as residential users, commercial users, public construction users, and / or industrial users.

[0170] Figure 3 A structural block diagram of a dynamic aggregation device for distributed resources in a transformer substation according to an embodiment of the present disclosure is shown. This device can be implemented as part or all of an electronic device through software, hardware, or a combination of both.

[0171] like Figure 3 As shown, the distributed resource dynamic aggregation device 300 includes a first determining module 310, a second determining module 320, a third determining module 330, a clustering module 340, and an aggregation module 350. Among them,

[0172] The first determining module 310 is configured to acquire a sample of the distributed resource and determine the degree matrix and adjacency matrix of the distributed resource based on the sample;

[0173] The second determining module 320 is configured to determine the normalized Laplacian matrix of the sample based on the degree matrix and the adjacency matrix;

[0174] The third determining module 330 is configured to determine the clustering matrix of the samples based on the standardized Laplace matrix;

[0175] The clustering module 340 is configured to use a genetic algorithm to cluster the clustering matrix to obtain multiple large-area clusters, each of which has a cluster center.

[0176] The aggregation module 350 is configured to use each cluster center as the regional control node of the transformer area, and to realize dynamic aggregation of distributed resources in the transformer area based on the regional control node.

[0177] In this embodiment of the disclosure, optimizing the clustering matrix using the first and second objective functions of the genetic algorithm to obtain preliminary clustering results includes: randomly selecting initial cluster centers p, and refining the elements α of the clustering matrix based on the minimum Euclidean distance criterion. iAllocate it to the corresponding cluster, where i is a positive integer, 0 < i ≤ N, and N is the total number of elements in the clustering matrix; recalculate the new clustering center p of the cluster to which the element ɑ has been allocated based on the first objective function i of the cluster new ; based on the new clustering center p new re-cluster the current cluster to obtain a new current cluster C new ; verify the distance between the new current cluster C new and other clusters based on the second objective function to ensure that the element ɑ i is allocated to the correct cluster.

[0178] In an embodiment of the present disclosure, the first objective function is a cluster validity objective function based on the within-cluster cohesion index, and the within-cluster cohesion index is determined by the mean square deviation E of the within-cluster distances where E j is the mean square deviation of the within-cluster distances of the j-th element in the current cluster, j is a positive integer, 1 ≤ j ≤ M, and M is the total number of elements in the current cluster.

[0179] In an embodiment of the present disclosure, the recalculating the new clustering center of the cluster to which the element ɑ i has been allocated based on the first objective function includes: recalculating the mean square deviation of the within-cluster distances of each element in the cluster to which the element ɑ i has been allocated based on the first objective function; determining the element with the smallest mean square deviation of the within-cluster distances as the new clustering center of the cluster.

[0180] In an embodiment of the present disclosure, the second objective function is a cluster validity objective function based on the between-cluster separation index, and the between-cluster separation index is determined by the between-cluster distance I, I = max‖C new - C m ‖, where m is a positive integer, m is less than or equal to the total number of clusters in the clustering matrix, and ‖C new - C m ‖ is to calculate the difference between the corresponding position elements in the new current cluster C new and the cluster C m .

[0181] In an embodiment of the present disclosure, the verifying the distance between the new current cluster C new and other clusters based on the second objective function to ensure that the element ɑ i is allocated to the correct cluster includes: calculating the between-cluster distance I between the new current cluster C new and other clusters, and when the between-cluster distance I is greater than the first threshold, determining that the element ɑ i is allocated to the correct cluster.

[0182] In this embodiment of the disclosure, the third objective function is an inter-cluster correlation objective function, determined by the inter-cluster covariance coefficient, which is:

[0183]

[0184] Where a is the cluster center of the first cluster A, and b is the cluster center of the second cluster B. It is the average value of all elements in the first cluster A. It is the average value of each element in the second cluster B.

[0185] In this embodiment of the disclosure, the step of optimizing the number of clusters in the preliminary clustering results using a third objective function to obtain multiple large-area aggregated clusters includes: calculating the inter-cluster covariance coefficient between each cluster, and merging clusters whose inter-cluster covariance coefficient is less than a second threshold; calculating the inter-cluster covariance coefficient of each merged cluster again until the inter-cluster covariance coefficient between any two clusters is greater than or equal to the second threshold; and determining that the final obtained clusters are the large-area aggregated clusters.

[0186] In this embodiment of the disclosure, the dynamic aggregation of distributed resources in the distribution area based on the regional control node includes: frequency modulation aggregation and / or peak modulation aggregation of distributed resources in the distribution area based on the regional control node.

[0187] In this embodiment of the disclosure, the regional control node is based on the response time T respond Response duration τ duration and / or responsive frequency ω respond The distributed resources of the aforementioned distribution area are dynamically aggregated; wherein, the response time T respond The response duration τ is the maximum time required for a resource to fully reach the output level required by the instruction after receiving it. duration After receiving a resource response command, the response must maintain the output state required by the command for the shortest possible time, wherein the responsive frequency ω respond This represents the number of times a resource can be accessed within a given period of time.

[0188] In this embodiment of the disclosure, the regional control node dynamically aggregates the distributed resources based on load frequency regulation performance indicators and / or load peak shaving indicators; wherein, the load frequency regulation performance indicator Fn = (1-t) / ( ... respond )ω respond , t respond =T respond / max(T 1，respond T 2，respond ,…,T N，respond );

[0189] The load peak shaving index Pn = (1-t)respond )τ duration .

[0190] In this embodiment of the disclosure, the distributed resources participating in frequency modulation need to meet the following constraints:

[0191] T n，respond ≤T F respond

[0192] τ n，duration ≥τ F duration

[0193] ω n，respond ≥ω F respond ;

[0194] Among them, T F respond τ is the upper limit of the acceptable response time for FM service. F duration For the duration of the response of frequency modulation service, ω F respond For the maximum responsive frequency; and / or

[0195] Distributed resources participating in peak shaving must meet the following constraints:

[0196] T n，respond ≤T P respond

[0197] τ n，duration ≥τ P duration ;

[0198] Among them, T P respond and τ P duration These are the maximum allowed response time and the minimum allowed duration for peak shaving services, respectively.

[0199] In this embodiment of the disclosure, obtaining samples of the distributed resources and determining the degree matrix D of the distributed resources based on the samples includes: obtaining load characteristic parameters of the distributed resources as the samples; determining characteristic parameters of the distributed resources based on the Euclidean distance of the load characteristic parameters; and constructing the degree matrix based on the characteristic parameters. Determining the adjacency matrix of the distributed resources includes: constructing the adjacency matrix based on a Gaussian kernel function. Determining the clustering matrix of the samples based on the normalized Laplacian matrix includes: the eigenvalues ​​and eigenvectors of the normalized Laplacian matrix; and constructing the clustering matrix using the eigenvectors corresponding to the k largest eigenvalues ​​as row vectors.

[0200] According to the technical solution provided in the embodiments of this disclosure, the dynamic aggregation of distributed resources in the distribution area is realized through the regional control node, which enables the intelligent aggregation of distributed resources in multiple distribution areas, improves the aggregation efficiency, and enhances the control capability of the distribution network.

[0201] This disclosure also discloses an electronic device. Figure 4 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0202] like Figure 4 As shown, the electronic device includes a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method for dynamic aggregation of distributed resources in a distribution area according to an embodiment of the present disclosure.

[0203] In this embodiment of the disclosure, the method for dynamic aggregation of distributed resources in a transformer substation includes: obtaining samples of the distributed resources; determining the degree matrix and adjacency matrix of the distributed resources based on the samples; determining the normalized Laplacian matrix of the samples based on the degree matrix and adjacency matrix; determining the clustering matrix of the samples based on the normalized Laplacian matrix; clustering the clustering matrix using a genetic algorithm to obtain multiple large-area clusters, each of the large-area clusters having a cluster center; using each cluster center as a large-area control node for the transformer substation, and realizing dynamic aggregation of distributed resources in the transformer substation based on the large-area control node.

[0204] This disclosure also provides a chip, which includes the above-mentioned dynamic aggregation device for distributed resources in the distribution area. The device can be implemented as part or all of the chip through software, hardware, or a combination of both.

[0205] Figure 5 A schematic diagram of the structure of a computer system suitable for implementing the method according to embodiments of the present disclosure is shown.

[0206] like Figure 5 As shown, the computer system includes a processing unit that can execute various methods described above based on a program stored in a read-only memory (ROM) or a program loaded from a storage portion into a random access memory (RAM). The RAM also stores various programs and data required for the operation of the computer system. The processing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0207] The following components are connected to the I / O interface: input sections including keyboards, mice, etc.; output sections including cathode ray tubes (CRTs), liquid crystal displays (LCDs), and speakers; storage sections including hard disks, etc.; and communication sections including network interface cards such as LAN cards and modems. The communication section performs communication processes via a network such as the Internet. Drives are also connected to the I / O interface as needed. Removable media, such as disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on the drive as needed so that computer programs read from them can be installed into the storage section as needed. The processing unit can be implemented as a CPU, GPU, TPU, FPGA, NPU, etc.

[0208] In particular, according to embodiments of this disclosure, the methods described above can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program containing program code for performing the methods described above. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium.

[0209] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0210] The units or modules described in the embodiments of this disclosure can be implemented in software or programmable hardware. The described units or modules can also be located in a processor, and the names of these units or modules do not necessarily constitute a limitation on the unit or module itself.

[0211] In another aspect, this disclosure also provides a computer-readable storage medium, which may be a computer-readable storage medium included in the electronic device or computer system described above; or it may be a standalone computer-readable storage medium not assembled into a device. The computer-readable storage medium stores one or more programs, which are used by one or more processors to perform the methods described in this disclosure.

[0212] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

Claims

1. A method for dynamic aggregation of distributed resources in a transformer substation, characterized in that, include: Obtain a sample of the distributed resource, and determine the degree matrix and adjacency matrix of the distributed resource based on the sample; The normalized Laplacian matrix of the sample is determined based on the degree matrix and the adjacency matrix. The clustering matrix of the samples is determined based on the standardized Laplacian matrix; The clustering matrix is ​​clustered using a genetic algorithm to obtain multiple large-area clusters, each of which has a cluster center. Each cluster center serves as the regional control node for the transformer area, and dynamic aggregation of distributed resources in the transformer area is realized based on the regional control nodes. The distributed resources in the transformer area communicate bidirectionally with the dispatch center, and the dispatch center is connected to the control bus and communicates bidirectionally. The control bus includes multiple regional control nodes to dynamically aggregate the distributed resources in the transformer area, thereby providing electricity resources to the electricity users. The method of using a genetic algorithm to cluster the clustering matrix to obtain multiple large-area clusters includes: Randomly select an initial clustering center p, and assign the element ɑ of the clustering matrix to the corresponding cluster based on the minimum Euclidean distance criterion, where i is a positive integer, 0 < i ≤ N, and N is the total number of elements in the clustering matrix; i i is a positive integer, 0 < i ≤ N, where N is the total number of elements in the clustering matrix; The element α was recalculated and redistributed based on the first objective function. i The new cluster center p of the cluster new ; Based on the new cluster center p new Re-cluster the current cluster to obtain a new current cluster C. new ; Verify the new current cluster C based on the second objective function. new Distance from other clusters to ensure element α i The elements in the clustering matrix are assigned to the correct clusters. The first clustering optimization is performed based on the first and second objective functions. Next, the number of clusters after the first clustering optimization is further optimized based on the third objective function, and the final cluster centers are generated. The inter-cluster covariance coefficients between each cluster are calculated based on the third objective function, and clusters with inter-cluster covariance coefficients less than the second threshold are merged; the inter-cluster covariance coefficients of the merged clusters are calculated again until the inter-cluster covariance coefficients between any two clusters are greater than or equal to the second threshold; the final clusters are determined to be the large-area aggregation clusters; The third objective function is the inter-cluster correlation objective function, which is determined by the inter-cluster covariance coefficient, and the inter-cluster covariance coefficient is: ； Where a is the cluster center of the first cluster A, and b is the cluster center of the second cluster B. It is the average value of all elements in the first cluster A. It is the average value of each element in the second cluster B.

2. The method according to claim 1, characterized in that, The first objective function is a cluster effectiveness objective function based on the intra-cluster cohesion index, wherein the intra-cluster cohesion index is determined by the mean square deviation E of the intra-cluster distance. , of which E j Let M be the mean squared deviation of the intra-cluster distance of the j-th element in the current cluster, where j is a positive integer, 1≤j≤M, and M is the total number of elements in the current cluster.

3. The method according to claim 2, characterized in that, The element α was recalculated and allocated based on the first objective function. i The new cluster centers of the clusters include: The element α was recalculated and redistributed based on the first objective function. i The mean square deviation of the intra-cluster distances of elements in a cluster; The element with the smallest mean square deviation of the intra-cluster distance is determined as the new cluster center of the cluster.

4. The method according to claim 3, characterized in that, The second objective function is a cluster effectiveness objective function based on the inter-cluster separation index, where the inter-cluster separation index is determined by the inter-cluster distance I, I = max(I / C). new -C m ǁ, where m is a positive integer, and m is less than or equal to the total number of clusters in the clustering matrix, ǁC new -C m ǁ is the calculation of the new current cluster C. new In and cluster C m The difference between the corresponding elements in the middle.

5. The method according to claim 4, characterized in that, The new current cluster C is verified based on the second objective function. new Distance from other clusters to ensure element α i Being assigned to the correct cluster, including: Calculate the new current cluster C new The element α is determined when the inter-cluster distance I with other clusters is greater than a first threshold. i It was assigned to the correct cluster.

6. The method according to claim 1, characterized in that, The dynamic aggregation of distributed resources in the distribution area based on the regional control node includes: Frequency modulation aggregation and / or peak modulation aggregation of distributed resources in the transformer area are realized based on the regional control node.

7. The method according to claim 6, characterized in that, The regional control node is based on the response time T. respond Response duration τ duration and / or responsive frequency ω respond Dynamically aggregate the distributed resources of the aforementioned transformer area; Wherein, the response time T respond The response duration τ is the maximum time required for a resource to fully reach the output level required by the instruction after receiving it. duration After receiving a resource response command, the response must maintain the output state required by the command for the shortest possible time, wherein the responsive frequency ω respond This represents the number of times a resource can be accessed within a given period of time.

8. The method according to claim 7, characterized in that, The regional control node dynamically aggregates the distributed resources based on load frequency regulation performance indicators and / or load peak regulation indicators; Wherein, the load frequency regulation performance index Fn = (1-t) respond )ω respond , t respond =T respond / max(T 1，respond ,T 2，respond ,…,T N，respond ); The load peak shaving index Pn = (1-t) respond )τ duration .

9. The method according to claim 8, characterized in that, Distributed resources participating in frequency modulation must meet the following constraints: T n，respond ≤T F respond t n，duration ≥τ F duration oh n，respond ≥ω F respond ; Among them, T F respond τ is the upper limit of the acceptable response time for FM service. F duration For the duration of the response of frequency modulation service, ω F respond For the maximum responsive frequency; and / or Distributed resources participating in peak shaving must meet the following constraints: T n，respond ≤T P respond t n，duration ≥τ P duration ; Among them, T P respond and τ P duration These are the maximum allowed response time and the minimum allowed duration for peak shaving services, respectively.

10. The method according to claim 1, characterized in that, The step of obtaining a sample of the distributed resource and determining the degree matrix D of the distributed resource based on the sample includes: The load characteristic parameters of the distributed resources are obtained as the sample; The characteristic parameters of the distributed resource are determined based on the Euclidean distance of the load characteristic parameters; The degree matrix is ​​constructed based on the feature parameters.

11. The method according to claim 10, characterized in that, Determining the adjacency matrix of the distributed resource includes: The adjacency matrix is ​​constructed based on the Gaussian kernel function.

12. The method according to claim 11, characterized in that, Determining the clustering matrix of the samples based on the standardized Laplacian matrix includes: The eigenvalues ​​and eigenvectors of the standardized Laplacian matrix; The clustering matrix is ​​constructed by using the eigenvectors corresponding to the top k largest eigenvalues ​​as row vectors.

13. A dynamic aggregation device for distributed resources in a transformer substation, characterized in that, include: The first determining module is configured to obtain a sample of the distributed resource and determine the degree matrix and adjacency matrix of the distributed resource based on the sample. The second determining module is configured to determine the normalized Laplacian matrix of the sample based on the degree matrix and the adjacency matrix; The third determining module is configured to determine the clustering matrix of the samples based on the standardized Laplacian matrix; The clustering module is configured to use a genetic algorithm to cluster the clustering matrix to obtain multiple large-area clusters, each of which has a cluster center. The aggregation module is configured to use each cluster center as the regional control node of the transformer area, and to realize the dynamic aggregation of distributed resources in the transformer area based on the regional control node; wherein, the distributed resources of the transformer area communicate bidirectionally with the dispatch center, the dispatch center is connected to the control bus and communicates bidirectionally, and the control bus includes multiple regional control nodes to dynamically aggregate the distributed resources of the transformer area, thereby providing electricity resources to the electricity users; The clustering module is implemented as follows: Randomly select an initial clustering center p, and assign the element ɑ of the clustering matrix to the corresponding cluster based on the minimum Euclidean distance criterion, where i is a positive integer, 0 < i ≤ N, and N is the total number of elements in the clustering matrix; i i is a positive integer, 0 < i ≤ N, and N is the total number of elements in the clustering matrix; The element α was recalculated and redistributed based on the first objective function. i The new cluster center p of the cluster new ; Based on the new cluster center p new Re-cluster the current cluster to obtain a new current cluster C. new ; Verify the new current cluster C based on the second objective function. new Distance from other clusters to ensure element α i The elements in the clustering matrix are assigned to the correct clusters. The first clustering optimization is performed based on the first and second objective functions. Next, the number of clusters after the first clustering optimization is further optimized based on the third objective function, and the final cluster centers are generated. The inter-cluster covariance coefficients between each cluster are calculated based on the third objective function, and clusters with inter-cluster covariance coefficients less than the second threshold are merged; the inter-cluster covariance coefficients of the merged clusters are calculated again until the inter-cluster covariance coefficients between any two clusters are greater than or equal to the second threshold; the final clusters are determined to be the large-area aggregation clusters; The third objective function is the inter-cluster correlation objective function, which is determined by the inter-cluster covariance coefficient, and the inter-cluster covariance coefficient is: ； Where a is the cluster center of the first cluster A, and b is the cluster center of the second cluster B. It is the average value of all elements in the first cluster A. It is the average value of each element in the second cluster B.

14. The apparatus according to claim 13, characterized in that, The first objective function is a cluster effectiveness objective function based on the intra-cluster cohesion index, wherein the intra-cluster cohesion index is determined by the mean square deviation E of the intra-cluster distance. , of which E j Let M be the mean squared deviation of the intra-cluster distance of the j-th element in the current cluster, where j is a positive integer, 1≤j≤M, and M is the total number of elements in the current cluster.

15. The apparatus according to claim 14, characterized in that, The element α was recalculated and allocated based on the first objective function. i The new cluster centers of the clusters include: The element α was recalculated and redistributed based on the first objective function. i The mean square deviation of the intra-cluster distances of elements in a cluster; The element with the smallest mean square deviation of the intra-cluster distance is determined as the new cluster center of the cluster.

16. The apparatus according to claim 15, characterized in that, The second objective function is a cluster effectiveness objective function based on the inter-cluster separation index, where the inter-cluster separation index is determined by the inter-cluster distance I, I = max(I / C). new -C m ǁ, where m is a positive integer, and m is less than or equal to the total number of clusters in the clustering matrix, ǁC new -C m ǁ is the calculation of the new current cluster C. new In and cluster C m The difference between the corresponding elements in the middle.

17. The apparatus according to claim 16, characterized in that, The new current cluster C is verified based on the second objective function. new Distance from other clusters to ensure element α i Being assigned to the correct cluster, including: Calculate the new current cluster C new The element α is determined when the inter-cluster distance I with other clusters is greater than a first threshold. i It was assigned to the correct cluster.

18. The apparatus according to claim 13, characterized in that, The dynamic aggregation of distributed resources in the distribution area based on the regional control node includes: Frequency modulation aggregation and / or peak modulation aggregation of distributed resources in the transformer area are realized based on the regional control node.

19. The apparatus according to claim 18, characterized in that, The regional control node is based on the response time T. respond Response duration τ duration and / or responsive frequency ω respond Dynamically aggregate the distributed resources of the aforementioned transformer area; Wherein, the response time T respond The response duration τ is the maximum time required for a resource to fully reach the output level required by the instruction after receiving it. duration After receiving a resource response command, the response must maintain the output state required by the command for the shortest possible time, wherein the responsive frequency ω respond This represents the number of times a resource can be accessed within a given period of time.

20. The apparatus according to claim 19, characterized in that, The regional control node dynamically aggregates the distributed resources based on load frequency regulation performance indicators and / or load peak regulation indicators; Wherein, the load frequency regulation performance index Fn = (1-t) respond )ω respond , t respond =T respond / max(T 1，respond ,T 2，respond ,…,T N，respond ); The load peak shaving index Pn = (1-t) respond )τ duration .

21. The apparatus according to claim 20, characterized in that, Distributed resources participating in frequency modulation must meet the following constraints: T n，respond ≤T F respond t n，duration ≥τ F duration oh n，respond ≥ω F respond ; Among them, T F respond τ is the upper limit of the acceptable response time for FM service. F duration For the duration of the response of frequency modulation service, ω F respond For the maximum responsive frequency; and / or Distributed resources participating in peak shaving must meet the following constraints: T n，respond ≤T P respond t n，duration ≥τ P duration ; Among them, T P respond and τ P duration These are the maximum allowed response time and the minimum allowed duration for peak shaving services, respectively.

22. The apparatus according to claim 13, characterized in that, The step of obtaining a sample of the distributed resource and determining the degree matrix D of the distributed resource based on the sample includes: The load characteristic parameters of the distributed resources are obtained as the sample; The characteristic parameters of the distributed resource are determined based on the Euclidean distance of the load characteristic parameters; The degree matrix is ​​constructed based on the feature parameters.

23. The apparatus according to claim 22, characterized in that, Determining the adjacency matrix of the distributed resource includes: The adjacency matrix is ​​constructed based on the Gaussian kernel function.

24. The apparatus according to claim 23, characterized in that, Determining the clustering matrix of the samples based on the standardized Laplacian matrix includes: The eigenvalues ​​and eigenvectors of the standardized Laplacian matrix; The clustering matrix is ​​constructed by using the eigenvectors corresponding to the top k largest eigenvalues ​​as row vectors.

25. An electronic device, characterized in that, The method includes a memory and a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the steps of the method according to any one of claims 1-12.

26. A chip, characterized in that, The chip includes the device as described in any one of claims 13-24.

27. A computer-readable storage medium storing computer instructions thereon, characterized in that, When executed by a processor, the computer instructions implement the steps of the method described in any one of claims 1-12.

Images